Research map / Detection and segmentation / Open-set and few-shot detection
Zero-Shot Semantic Segmentation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| generate synthetic unseen-class features with GMMN and retrain classifier instead of embedding-regression baseline | DeepLabv3+ · Pascal-VOC 2 unseen | hIoU: 6.1 -> 47.5 | Helped |
| generate synthetic unseen-class features instead of embedding-regression baseline | DeepLabv3+ · Pascal-VOC 10 unseen | hIoU: 3.6 -> 23.6 | Helped |
| add graph-context encoding to the generator (ablation) | ZS3Net · Pascal-Context 2 unseen | hIoU: 28.4 -> 34.8 | Helped |
| add graph-context encoding to the generator (ablation) | ZS3Net · Pascal-Context 10 unseen | hIoU: 15.8 -> 17.8 | Helped |
| add self-training on pseudo-labeled unseen pixels (ZS5Net) | ZS3Net · Pascal-VOC 2 unseen | hIoU: 47.5 -> 75.7 | Helped |
| add self-training on pseudo-labeled unseen pixels (ZS5Net) | ZS3Net · Pascal-VOC 10 unseen | hIoU: 23.6 -> 46.7 | Helped |
| evaluate with generalized ZSL protocol rather than vanilla (seen scores included) | Baseline (DeViSe) · Pascal-VOC 10 unseen | unseen mIoU: 41.7 -> 1.9 | Hurt |
| use GAN instead of GMMN as feature generator (ablation) | ZS3Net · Pascal-VOC | mIoU | No clear effect |
Closest papers by meaning
- A Simple Framework for Open-Vocabulary Zero-Shot Segmentation (ICLR 2025 · similarity 0.89)
- S3A: Towards Realistic Zero-Shot Classification via Self Structural Semantic Alignment (AAAI 2024 · similarity 0.89)
- A Surprisingly Simple Approach to Generalized Few-Shot Semantic Segmentation (NeurIPS 2024 · similarity 0.89)
- DiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut (NeurIPS 2024 · similarity 0.89)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.89)
- TPR: Topology-Preserving Reservoirs for Generalized Zero-Shot Learning (NeurIPS 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.